Ask Your Question

What are the Tag Results for all Queries since 2012?

asked 2020-08-24 17:15:17 +0100

Alex Kemp gravatar image

updated 2020-09-20 21:54:20 +0100


This site has been operating since 2012 and, as I ask this question, contains 42,100 queries. Most of those questions are about LibreOffice, and every question is required to have at least one (very mysterious) Tag attached.

Yesterday I collected every tag for every question ever asked, put them into Calc and did an analysis, with 3 charts made of the results (first below is the image of the chart for the spread of OS reported, then the 3 charts that are currently available):


  1. LibreOffice Package Queries (36,549 total)
  2. LibreOffice Version Queries (7,393 total)
  3. Operating System Queries (7,523 total)

Now, as you can tell, those charts need improvement. Nevertheless, I hear you say, where can we get the results? Well, the ODS is linked above (tags-numbers-charts.ods).

How did you do it, Alex?

  1. Collect the Tags using a wget script (creates 165 text-files):

    for a in {1..165};    
        get="sort\=used\&page=$a"; wget -O tags-$a \   $get;    
  2. Extract each tag+number from each file as above into a .csv file:

    for a in {1..165};
        fgrep -B1 '×' tags-$a | \
        awk '{ \
            if(index($1,"data-tag-name")) \
                { printf $1" "; getline; print $3 } \
        }' | \
        sed 's/^data-tag-name="\([^"]*\)".*215;\([,0-9]*\).*/\1 \2/' - >> tags.csv;
  3. Use the CSV import routine to create a list of all the tags together with it's number in the aslibo DB:
    image description
    That creates tags.csv (same as the original file) and needs to be saved as a standard ODS file (available at top as tags-list.ods).

  4. That list of tags needs de-duping.
    I installed remove-duplicates-fast. It removed an astonishing 1,173 dupes.
    image description  
  5. Classify & name the tags + produce the charts.
    ...and there you have it. Hopefully interesting.

If this helps then please tick the answer (✔)
...and/or show you like it with an uptick ()

(edit: created proper code sections (originals are each all on one line))

edit retag flag offensive close merge delete


Hit a still-current bug#112102 which prevents labels being placed outside the graphic in DoughNut charts (like the one shown above) because the Placement option is greyed out. Really annoying!

Switching to another type of chart (other than Donut) will allow Placement to become available, and that option may even carry over if the chart-option is switched back. However, a multi-series DoughNut chart (like the OS chart above) still has a greyed-out value + default value for that option, even after trying the work-around. The bug-report was opened 3 years ago & is probably much earlier, since the Version affected is shown to be "Inherited From OOo".

Alex Kemp gravatar imageAlex Kemp ( 2020-08-24 17:48:28 +0100 )edit

1 Answer

Sort by » oldest newest most voted

answered 2020-08-27 10:06:44 +0100

Alex Kemp gravatar image

OS frequency:

Windows + Linux are approaching equal-numbers on the desktop:

  1. Windows: 44%
  2. Linux: 35%
  3. MacOS: 15%

The tag-numbers include results for Android, iOS, iPad & Java which skews the figures above a little + the top-3 do include some numbers for servers within what are otherwise desktop machines. Nevertheless, those other numbers are tiny & can be ignored.

I first met Linux in 1998 whilst supporting users of the UK's 1st national ISP. That was through one of my colleagues that made use of it on a desktop machine. A few months later I set up a website under Linux on a remote server. Linux was well-known on servers at that time, and almost unknown on the desktop (at a guess, less than 1%). Boy, has that changed.

edit flag offensive delete link more
Login/Signup to Answer

Question Tools

1 follower


Asked: 2020-08-24 17:15:17 +0100

Seen: 90 times

Last updated: Sep 20 '20